ROL:       true,
	ROR:       true,
	SAR:       true,
	SBB:       true,
	SHL:       true,
	SHLD:      true,
	SHR:       true,
	SHRD:      true,
	SUB:       true,
	TEST:      true,
	XADD:      true,
	XCHG:      true,
	XOR:       true,
}

var plan9Reg = [...]string{
	AL:   "AL",
	CL:   "CL",
	BL:   "BL",
	DL:   "DL",
	AH:   "AH",
	CH:   "CH",
	BH:   "BH",

func semrelease(addr *uint32) {
	semrelease1(addr, false, 0)
}

func semrelease1(addr *uint32, handoff bool, skipframes int) {
	root := semtable.rootFor(addr)
	atomic.Xadd(addr, 1)

	// Easy case: no waiters?
	// This check must happen after the xadd, to avoid a missed wakeup
	// (see loop in semacquire).
	if root.nwait.Load() == 0 {
		return
	}

	// Harder case: search for a waiter and wake it.

		ret := lwp_park(_CLOCK_MONOTONIC, _TIMER_RELTIME, tsp, 0, unsafe.Pointer(&gp.m.waitsemacount), nil)
		if ret == _ETIMEDOUT {
			return -1
		}
	}
}

//go:nosplit
func semawakeup(mp *m) {
	atomic.Xadd(&mp.waitsemacount, 1)
	// From NetBSD's _lwp_unpark(2) manual:
	// "If the target LWP is not currently waiting, it will return
	// immediately upon the next call to _lwp_park()."

				job := atomic.Xadd(&work.markrootNext, +1) - 1
				if job < work.markrootJobs {
					workFlushed += markroot(gcw, job, false)
					continue
				}
			}
			// No heap or root jobs.
			break
		}

		scanobject(b, gcw)

		// Flush background scan work credit.
		if gcw.heapScanWork >= gcCreditSlack {
			gcController.heapScanWork.Add(gcw.heapScanWork)

			// External linker will do this relocation.
			return true
		}

		// Internal linking, build a PLT entry and change the relocation
		// target to that entry.
		if r.Add() != 0 {
			ldr.Errorf(s, "PLT call with non-zero addend (%v)", r.Add())
		}
		addpltsym(target, ldr, syms, targ)
		su := ldr.MakeSymbolUpdater(s)
		su.SetRelocSym(rIdx, syms.PLT)
		su.SetRelocAdd(rIdx, int64(ldr.SymPlt(targ)))
		return true
	}

		su := ldr.MakeSymbolUpdater(s)
		su.SetRelocType(rIdx, objabi.R_PCREL)
		ldr.SetRelocVariant(s, rIdx, sym.RV_390_DBL)
		su.SetRelocAdd(rIdx, r.Add()+int64(r.Siz()))
		if targType == sym.SDYNIMPORT {
			addpltsym(target, ldr, syms, targ)
			r.SetSym(syms.PLT)
			su.SetRelocAdd(rIdx, r.Add()+int64(ldr.SymPlt(targ)))
		}
		return true

	case objabi.ElfRelocOffset + objabi.RelocType(elf.R_390_PLT32),

//
//go:nosplit
func (u *Uint32) Or(value uint32) {
	Or(&u.value, value)
}

// Add adds delta to u atomically, returning
// the new updated value.
//
// This operation wraps around in the usual
// two's-complement way.
//
//go:nosplit
func (u *Uint32) Add(delta int32) uint32 {
	return Xadd(&u.value, delta)
}

// Uint64 is an atomically accessed uint64 value.
//

		// Flush the write barrier buffer, since this may add
		// work to the gcWork.
		wbBufFlush1(pp)

		// Flush the gcWork, since this may create global work
		// and set the flushedWork flag.
		//
		// TODO(austin): Break up these workbufs to
		// better distribute work.
		pp.gcw.dispose()
		// Collect the flushedWork flag.
		if pp.gcw.flushedWork {
			atomic.Xadd(&gcMarkDoneFlushed, 1)
			pp.gcw.flushedWork = false

					finptrmask[i] = finalizer1[i%len(finalizer1)]
				}
			}
		}
		block := finc
		finc = block.next
		block.next = finq
		finq = block
	}
	f := &finq.fin[finq.cnt]
	atomic.Xadd(&finq.cnt, +1) // Sync with markroots
	f.fn = fn
	f.nret = nret
	f.fint = fint
	f.ot = ot
	f.arg = p
	unlock(&finlock)
	fingStatus.Or(fingWake)
}

//go:nowritebarrier

		su.SetRelocAdd(rIdx, r.Add()+8)
		return true

	case objabi.ElfRelocOffset + objabi.RelocType(elf.R_X86_64_PLT32):
		su := ldr.MakeSymbolUpdater(s)
		su.SetRelocType(rIdx, objabi.R_PCREL)
		su.SetRelocAdd(rIdx, r.Add()+4)
		if targType == sym.SDYNIMPORT {
			addpltsym(target, ldr, syms, targ)
			su.SetRelocSym(rIdx, syms.PLT)
			su.SetRelocAdd(rIdx, r.Add()+int64(ldr.SymPlt(targ)))
		}